Cannot print a model with imports?

[kerimoyle]

Say I have two models, one of which imports the other (see below). According to 2.15.1 and 3.4.2.2, I'm allowed to have a map_variables item that points to an imported component. How can I get this output using the API?

• if I serialise the model without flattening, there can be no equivalences stored as the variables are not instantiated;
• if I serialise the model after flattening, then it's no longer an import situation ... ?

<?xml version="1.0" encoding="UTF-8"?>
<model xmlns="http://www.cellml.org/cellml/2.0#" name="import_example">
  <import xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="importMe.cellml">
    <component component_ref="importedComponent" name="importMe"/>
  </import>
  <component name="localComponent">
    <variable name="x" units="dimensionless" interface="public_and_private"/>
  </component>
  <connection component_1="localComponent" component_2="importedComponent">
    <map_variables variable_1="x" variable_2="x"/> ///**** NOTE THIS BIT!!
  </connection>
</model>

Inside importMe.cellml:

<?xml version="1.0" encoding="UTF-8"?>
<model xmlns="http://www.cellml.org/cellml/2.0#" name="import_me">
  <component name="importMe">
    <variable name="x" units="dimensionless" interface="public_and_private"/>
  </component>
</model>

[nickerso]

Maybe throw together a test which demonstrates this behaviour, that would be the easiest way to look at this. Cause as it stands this sounds like a huge bug.

[hsorby]

small bug

[kerimoyle]

I've written two small tests in #591 ... the second one (in variable tests) won't compile, but it shows what I'd expect to be able to do?

[nickerso]

So as we remember how this was designed to work, there is actually no bug (see discussion on #591). But there could be new API to make the solution more convenient.

[kerimoyle]

So as we remember how this was designed to work, there is actually no bug (see discussion on #591). But there could be new API to make the solution more convenient.

Nope, sorry @nickerso. There is a bug in that the variable equivalences are not being printed (see first test in PR) How the equivalence is established in the first place is a different thing ...

[nickerso]

is this the issue reported in #594? or do you mean something else?

[kerimoyle]

Something else - the original reason for this issue. The printer is not printing equivalences between variables when one (or more) of their components is imported. ie: the first test in #591